硅酸盐学报

直拉单晶硅氧浓度控制技术研究进展

PDF下载

刘文凯 ¹ , 刘赟 ¹ , 薛忠营 ¹^,² , 魏星 ¹^,²

硅酸盐学报 | 大尺寸高品质人工晶体制备与应用专题—综合评述 2025,53(12): 3494-3505

收起

硅酸盐学报 |大尺寸高品质人工晶体制备与应用专题—综合评述 2025 , 53 (12) : 3494 -3505

直拉单晶硅氧浓度控制技术研究进展

全屏

刘文凯¹, 刘赟¹, 薛忠营¹^,², 魏星¹^,²

作者信息

^1．集成电路材料国家重点实验室，中国科学院上海微系统与信息技术研究所上海 200050

^2．上海市高端硅基材料重点实验室，上海 200050

通讯作者:

魏星（1981—），男，博士，研究员。

作者简介:

刘文凯（1997—），男，博士研究生。

Research Progress on Oxygen Concentration Control Technology of Czochralski Monocrystalline Silicon

Wenkai LIU¹, Yun LIU¹, Zhongying XUE¹^,², Xing WEI¹^,²

Affiliations

^1.State Key Laboratory of Materials for integrated Circuits, Shanghai institute of Microsystem and Information Technology, Shanghai 200050, China

^2.Shanghai Key Laboratory of Advanced Silicon-Based Materials, Shanghai 200050, China

Corresponding Author:

WEI Xing (1981-), male, Ph.D., Professor. E-mail: xwei@mail.sim.ac.cn

LIU Wenkai (1997-), male, Doctoral candidate. E-mail: liuwk@mail.sim.ac.cn

出版时间: 2025-12-20 doi: 10.14062/j.issn.0454-5648.20250536

文章导航

摘要

收起

直拉单晶硅是当前集成电路制造的主流衬底材料。在直拉单晶硅生长过程中，石英坩埚的溶解会引入氧杂质，而单晶硅中的氧浓度直接影响衬底的电阻率、机械强度、金属吸杂能力和载流子寿命，因此，深入研究掌握直拉单晶硅氧输运的物理机制并实现晶体氧浓度的精确控制至关重要。本文首先介绍了直拉单晶硅生长的氧输运物理机制，并从坩埚溶解控制、熔体流动控制和自由液面挥发控制3个方面综述了氧浓度控制技术的研究进展，最后展望了氧控制技术的未来研究方向。

关键词

直拉单晶硅 / 氧输运 / 坩埚溶解 / 熔体流动 / 自由液面挥发

Abstract

收起

Czochralski (CZ) monocrystalline silicon is a dominant substrate material for integrated circuit manufacturing, accounting for over 90% of global silicon wafer production. During CZ silicon growth, oxygen impurities dissolve from quartz crucibles (SiO₂) into silicon melt at sustained high temperatures, ultimately incorporating into the crystal lattice. Oxygen in silicon affects essential substrate properties, including resistivity, mechanical strength, metal gettering capability, and carrier lifetime. Different applications impose divergent oxygen specifications. For instance, solar cells require concentrations of <7×10¹⁷ atoms·cm^-3 to minimize minority carrier lifetime degradation. Insulate Gate Bipolar Transistor (IGBT) power devices demand ultra-low oxygen (i.e., <2.5×10¹⁷ atoms·cm^-3) to suppress leakage current and breakdown voltage fluctuations. Furthermore, 3D-NAND flash memory necessitates a precise oxygen precipitate control to balance thermo-mechanical stress in multi-layer architectures. It is thus paramount for device performance and manufacturing yield to understand oxygen transport physics and achieve precise concentration control.

Oxygen transport in Cz silicon growth involves four primary mechanisms, i.e., 1) Crucible Dissolution: Oxygen dissolves from the quartz crucible wall into the silicon melt via the reaction (SiO₂(s) ⇌ Si(l)+2O(l)). The temperature dependence of oxygen solubility remains a subject of debate, 2) Melt Transport: Dissolved oxygen moves through the melt via convection and diffusion (governed by convection-diffusion equations), heavily influenced by melt flow dynamics, 3) Free Surface Volatilization: Oxygen evaporates at the melt surface as SiO gas (Si(l)+O(l) ⇌ SiO(g)), affected by argon ambient pressure and flow rate, and 4) Crystallization Segregation: The amount of oxygen that incorporates into crystal at the solid-liquid interface is affected by the solubility difference of oxygen between the silicon melt and the crystal. The complex, highly nonlinear interplay between these mechanisms presents a core scientific challenge. As segregation effects are dominated by the intrinsic characteristics of oxygen and silicon atoms and are difficult to control, the related research advancements involve.

1) Crucible Dissolution Control: Conventional approaches focus on thermal field optimization. Implementing dual-heater systems improves power distribution, compared to single heaters, effectively lowering localized crucible temperatures and subsequently reducing crystal oxygen content. Alternative strategies involve material modifications, such as applying Si₃N₄ coatings to quartz crucibles or using graphite crucibles with similar coatings, which can significantly suppress oxygen dissolution. However, these material substitutions introduce substantial trade-offs, including elevated carbon contamination (i.e., >5×10¹⁷ atoms·cm^-3) and nitrogen doping (i.e., >3×10¹⁵ atoms·cm^-3), requiring a further evaluation.

2) Melt Flow Control: Melt convection plays a pivotal role in oxygen transport. Precise manipulation of crystal and crucible rotation rates and directions impacts flow patterns and oxygen distribution, though identifying optimal settings becomes a challenge due to complex, field-dependent behavior. Magnetic fields are essential for stabilizing turbulent flow in large melts and reducing oxygen concentration consequently. Transverse Magnetic Fields (TMF) provide a strong turbulence suppression (i.e., 0.5 T) but often induce radial oxygen non-uniformity. Conversely, Cusp Magnetic Fields (CMF) offer a better radial uniformity but with lower field strengths (i.e., 0.1 T). The impact of Zero-Gauss plane (ZGP) position and coil current ratio varies considerably, indicating context-specific dependence. Innovations like quadrupole magnetic fields show a promise for unifying transverse-field stability with an enhanced radial homogeneity.

3) Free surface Volatilization Control: This strategy is to control oxygen concentration via SiO evaporation via manipulating the ambient argon environment. Lower argon pressure and higher argon flow rates generally promote SiO volatilization kinetics and reduce crystal oxygen incorporation, as reduced gas-phase partial pressure facilitates oxygen evaporation at the melt-gas interface. This relationship is consistently determined in both experimental measurements and theoretical models. Complementing gas parameter adjustments, heat shield design optimization offers a secondary control via altering the gas flow geometry above the melt. Specifically, tailored shield structures minimize argon recirculation zones that otherwise trap SiO vapor, thereby reducing local saturation and enhancing overall volatilization efficiency. However, shield modifications are less common in semiconductor-grade silicon production because they affect thermal gradients near the crystal, potentially disrupting the critical v/G ratio required for defect-free crystal growth. Furthermore, volatilization efficiency is not solely governed by gas dynamics. Complex interface phenomena on the melt surface (i.e., Marangoni convection and argon shear stress effects) also significantly affect oxygen evaporation rates. The interaction between these melt-surface flows and gas-phase mass transfer remains inadequately quantified, representing an important area for future fundamental research to fully exploit volatilization-based oxygen control strategies.

Summary and prospects

This review analyzes critical oxygen control strategies in CZ silicon growth, essential for achieving substrate specifications across diverse applications. The primary approaches focus on reducing oxygen dissolution through thermal field design and crucible modification, optimizing melt convection and oxygen transport via rotation and magnetic fields, and enhancing free surface volatilization by controlling argon atmosphere.

Some challenges persist, requiring a further research to clarify fundamental mechanisms. Key uncertainties include the temperature dependence of oxygen solubility and the precise mechanism relating crucible temperature reduction to lower crystal oxygen levels (primarily due to reduced dissolution or altered flow patterns). Understanding the complex coupling among detailed magnetic field distributions, resulting melt flow patterns and oxygen transport also requires a further investigation. Quantifying the influence of argon shear stress interacting with Marangoni forces on the melt surface is equally essential for refining volatilization models.

Furthermore, complicating these challenges is the multi-variable, strongly coupled nature of the CZ system, which makes conventional trial-and-error optimization approaches time-consuming and costly. A future research should therefore prioritize the developing intelligent control systems that leverage artificial intelligence (AI) and machine learning for data-driven modeling, prediction, and efficient multi-parameter optimization.

Key words

Czochralski monocrystalline silicon / oxygen transport / crucible dissolution / melt flow / melt free surface volatilization

引用本文

刘文凯, 刘赟, 薛忠营, 魏星. 直拉单晶硅氧浓度控制技术研究进展. 硅酸盐学报, 2025 , 53 (12) : 3494 -3505 . DOI: 10.14062/j.issn.0454-5648.20250536

Wenkai LIU, Yun LIU, Zhongying XUE, Xing WEI. Research Progress on Oxygen Concentration Control Technology of Czochralski Monocrystalline Silicon[J]. Journal of the Chinese Ceramic Society, 2025 , 53 (12) : 3494 -3505 . DOI: 10.14062/j.issn.0454-5648.20250536

基金

收起

中国科学院基础与交叉前沿科研先导专项（B类先导专项）(XDB0670102)
国家自然科学基金(62074152; 62304232; 62304233)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

2025年第53卷第12期

PDF下载

引用本文

BibTeX

文章信息

doi: 10.14062/j.issn.0454-5648.20250536

接收时间：2025-07-08
首发时间：2026-10-21
出版时间：2025-12-20

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2025-07-08
修回日期：2025-07-29

基金

中国科学院基础与交叉前沿科研先导专项（B类先导专项）(XDB0670102)

国家自然科学基金(62074152; 62304232; 62304233)

作者信息

^1．集成电路材料国家重点实验室，中国科学院上海微系统与信息技术研究所上海 200050

^2．上海市高端硅基材料重点实验室，上海 200050

通讯作者:

魏星（1981—），男，博士，研究员。

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/gsyxb/CN/10.14062/j.issn.0454-5648.20250536

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT